1. Field of the Invention
The present invention relates to devices for transmitting sound, specifically to speaker systems that utilize a cone reflector to reflect sound waves in a pattern resulting from the shape of the cone reflector.
2. Background Information
All speakers have a roll off in their frequency response as the speaker cabinet face becomes small relative to the wavelength of the sound being produced. This roll off of radiation efficiency is called diffraction loss. Diffraction loss adversely effects the low end frequency response of the speakers, leaving them sounding tinny. The higher sounds, having smaller wavelengths, are louder than lower sounds.
The transition frequency for diffraction loss occurs at a frequency whose one half wavelength occurs at the shortest width of the cabinet face. Above the transition frequency the speaker driver radiates as a hemisphere or 2 pi radians. Below the transition frequency the speaker driver radiates as a full sphere or 4 pi radians. The difference between these two different radiation patterns is 6 decibel of frontal lobe directivity gain for hemispherical radiation above the transition frequency. The cabinet face can be thought of as a 180 degree horn with the cutoff frequency at the width of the cabinet face. The total sound power into the room is the same above and below the transition frequency. Therefore, the problem exists that on axis frequency response is very different from off axis frequency response. This would occur even if the speaker driver was perfect. Real voices, instruments and microphones do not have this problem because they are acoustically small relative to the frequencies they produce or measure.
A conventional mini speaker may have a cabinet face dimension of 4 inches by 8 inches. These dimensions correspond to one half wavelength frequencies of 1695 Hertz and 847 Hertz. This results in a 6 decibel frequency step right in the middle of the voice and most instruments.
The diffraction loss effect could be corrected in a conventional speaker by adding 6 dB of electronic equalization. However, 6 dB of boost requires four times the amplifier power. In addition, a 6 dB boost would require a doubling of speaker diaphragm travel which would also raise Frequency Modulation Distortion by 6 dB. Other 2nd and 3rd harmonic distortions related to nonlinear BL product versus voice coil position would also be created. There would also be some power compression resulting in speaker parameter and frequency response changes. The cone area could be doubled to bring the diaphragm travel back to unity, but the extra mass would reduce height frequency extension and the larger diameter would make high frequencies more directional.
Another problem with conventional speakers is near field reflection. Near field reflection introduces distortion due to the small amount of delay time in the reflected sound. In research by Don Davis it is suggested that the minimum reflection time delay should be 10 msec (or approximately 8.85 feet path length) to avoid imaging problems. In a conventional speaker system a tweeter, or high frequency radiator, will be mounted some distance above the surface the speaker system is sitting on. When listening to the speaker there are two arrival times for the sound coming from the tweeter. The first arrival time is from the direct radiation of the tweeter to the ear and the second arrival time is from the reflection of the tweeter sound from the surface the speaker system is sitting on. The short delay time of the reflected sound causes "time smearing" of high frequencies which significantly reduces intelligibility and imaging of the sound. In addition, there is a dip in the frequency response due to the reflected wave being out of phase with the direct radiated wave. If a tweeter were 6 inches above a table top, with the listening ear 15 inches above the table top and 24 inches away from the speaker there will be an audible depression in the frequency response of the speaker centering around 1970 Hz. This corresponds to a difference in path length of 6.9894 inches resulting in a time delay of 515 micro-seconds.
An additional source of distortion occurs with ceiling mounted speakers when reflections of the sound waves arrive at the ear as a mono signal. Ceiling speakers have a relatively short time delay between the direct radiation from the ceiling and the reflected radiation from a desk top. Path length differences of 30 inches result in a 2190 micro-second delay which yields a frequency depression around 452 Hz. This tends to blur consonants of speech thereby reducing intelligibility.
There are two schools of thought on how to control the audibility of reflections. The first and most widely used in recording studios is the LEDE or Live End Dead End. This approach uses directional horn speakers with extensive room acoustic treatment. A second approach, which has been pursued for home reproduction, uses the principle of multiple diffuse reflections to mask and prevent any singular or speaker-based loud reflections from becoming clearly audible.
Basically six methods of achieving multiple diffuse reflections exist in the marketplace. The most widely known of the techniques is the BOSE approach. In the BOSE system discrete drivers are pointed in different directions. Although the result approximates uniform dispersion, due to its discrete nature the radiation pattern of these speakers is not continuous over 360 degrees. There is, therefore, severe comb filtering effects in the horizontal plane due to the individual drivers interacting. Further, the multiple drivers used do not maintain time alignment across the frequency band. This also disrupts the frequency balance and imaging through the crossover region. The reflected frequency balance can therefore be so distorted that conventional speakers will usually sound better than these designs.
The second most widely known technique is the Di-Polar approach used in electrostatic and ribbon speakers like Magnaplaner. This design uses the speakers without a rear enclosure or "open back". This design cancels all sound radiation to the sides, and rear sound is out of phase with the front sound. At low frequencies this cancellation drops the bass volume below perceptibility. Traditionally wide diaphragms are used. These types of diaphragms have high directivity change versus frequency. Thus, this radiation pattern does not create diffuse room reflections with even frequency balance. There is only one reflection off the back wall so it fails to mask room echoes. Di-Polar speakers also require ten times the air volume displacement of a box speaker for a given loudness due to the front/rear cancellations. They must therefore be very large to get significant volume output.
The third most widely known technique is Bi-Polar radiation. This approach is essentially placing two conventional speakers back to back with specific crossover changes. The design was first popularized by Mirage based on research by the Canadian National Research Council. Multiple drivers are placed on the front and back of the cabinet and operated in phase. The multiple diaphragms and shape of the cabinets cause very nonlinear frequency balance to the sides of the speakers. The rear speakers direct path sound wraps around the cabinet and combines with the front sound. The result is a large bump in frequency balance. The vertical offset of the drivers also causes vertical lobing error problems.
The fourth most widely known approach uses a reflector cone of some geometry. Reflector cones to date have been designed with curved sides used to encourage laminar air flow and to disperse the sound in the vertical plane. With traditional types of cone geometry approximately 25 percent of the sound is reflected back into the speaker. In addition, since the curved upper cone geometry includes included angles of less than 90 degrees in most designs, high frequency energy is directed below the speaker's horizontal plane. This results in secondary near field reflections. If the curved upper cone geometry includes curves of too small a diameter having included angles of greater than 90 degrees sounds are directed back into the speaker creating secondary reflections with severe frequency modulation distortion and comb filtering.
In addition, the curved reflector cones tend to reflect too much energy toward the ceiling. For instance, if the curved reflector cone includes included angles of greater than 135 degrees, energy is directed at an angle greater than 45 degrees above the horizontal plane. The energy at this angle tends to reflect off the ceiling before being heard by the listener, creating a reflection problem. In addition, the curved surface causes multiple phase delays in the high frequency which smears the transient response degrading high frequency output and reducing imaging.
The fifth type of 360 degree radiation speaker uses the rear radiation of a very special full range speaker driver constructed with its reflector cone having a very narrow included angle of only 45 degrees. This is the famous Lincoln Walsh design manufactured by OHM acoustics. This floor standing system mounts the driver on top of a box at ear level with the front of the driver facing down into the box. The listener listens to the back side of the moving speaker cone which sends sound 360 degrees in the horizontal plane except for high frequency which is absorbed in the rear 180 degrees with acoustic treatment. This design has some diffraction loss but its diffraction loss is partially compensated by the reduced high frequency efficiency of the full range driver. Less expensive designs by OHM use one separate conventional dome tweeter facing forward crossing over to a conventional bass/midrange driver placed in the Walsh configuration. In this two driver arrangement the directivity above and below the crossover is radically different.
The sixth type of 360 degree radiation speaker consists of pulsating cylinders stacked one above the other like in the German MBL speakers. They do have 360 degree radiation with identical frequency and volume. However, the vertical offset of the treble, midrange and bass drivers does cause significant horizontal lobing errors in the frequency response. There is also diffraction loss in this design.
It is clear that the speaker designs used to date do not overcome the above problems to provide identical frequency balance and volume in all directions of the horizontal plane. What is needed is a system and method of radiating sound energy uniformly and with identical frequency balance in all directions of the horizontal plane.